Performance tip No. 3: Let auld acquaintance be forgot
Many businesses cling desperately to elderly application platforms, leaving IT saddled with the high-cost, resource-intensive task of shoehorning old platforms into new infrastructures. This is how you wind up with a brand-new VMware vSphere architecture running a handful of Windows NT4 boxes.
Refusing to let go of the past often results in increased costs, downtime, and fragility of core business systems. Instead of holding meeting after meeting to figure out how to get a 10-year-old accounting package transferred to a new infrastructure, launch it into orbit and migrate to something new. The upfront costs may be more, but they will pale against the long-term costs you'll incur by not severing these ties.
This is a personnel issue as much as it is a technical one. There are always those in IT shops who see everything through the prism of their favored technology, facts be damned. It's not always easy to shepherd these folks through the dark and stormy night of new technology, but remember, hanging on to fixed-purpose IT admins can be as detrimental as hanging on to elderly technology.
Performance tip No. 4: Build a lab
There's no excuse. For the cost of a single server, you can build a monster IT test lab. A cheap, dual-CPU, 12-core AMD Istanbul-based 1U server can run several dozen virtual machines in a test scenario for about $1,500. Using VMware Server on Linux or VMware ESXi, you can avoid software licensing fees, while maintaining a perfectly valid platform for testing anything, from software upgrades to new packages, new operating systems, or even network architectures.
Combine a virtualized server lab with tools such as GNS3, and you can build and test just about any planned network or system infrastructure you want. There's no easier way to determine where resource bottlenecks reside than in a test bed, and if that test bed is as easily constructed as it is in a virtual lab, there's no reason not to find them. Moreover, with a virtual lab, you can find the sweet spot for certain servers, including how much RAM and CPU resources they'll need to function under expected (and unexpected) loads, thereby ensuring you waste fewer resources.
Performance tip No. 5: Watch everything
Network and system monitoring is the granddaddy of bottleneck diagnostics. When users complain that the network is slow, the network usually has nothing to do with it. But unless you have the facilities to show exactly where the problem resides, you're left hunting around in the dark for the solution.
Whether you prefer proprietary or open source tools, there's a myriad of options available to monitor everything from network latency and throughput to RAM and CPU utilization, to SAN performance and disk queue lengths -- you name it.
If it exists, it can be monitored. If it can be monitored, it can be graphed. And if it can be graphed, there's a very good chance that a simple perusal of the resulting graph can lead you in the right direction, greatly accelerating the problem-detection portion of any troubleshooting effort.